For years, plant personnel have relied upon a variety of traditional considerations—pressure rating, pressure drop, flowing medium, temperature and cost—to select a control valve. However, things have changed dramatically in the last decade. Advances in valve design, changes in process economics and a greater understanding of process controls have combined to render many of those traditional considerations significantly less relevant.
While the traditional considerations are important, they focus only on the “static” performance of the valve. In essence, they are “bench” measurements that say relatively nothing about how the valve will perform under actual operating conditions.
Conventional wisdom assumes that taking care of static considerations will provide good valve performance, thereby yielding good loop performance. We now know that this isn’t always the case.
Thousands of performance audits conducted by independent consultants and manufacturers have proven that as many as 50 percent of the valves in service, many of which were selected by conventional considerations, detract from the efforts to optimize control loop performance. Follow-up research shows that the valve’s dynamic characteristics play an important role in reducing process variability, which is part and parcel to process optimization. (Process variability is a measure of how closely the system can maintain the process variable to the set point despite random disturbances.)
However, note that both Valves B and C can achieve significantly greater reductions in process variability. With moderate loop tuning, the difference can be more than one percent. In many critical processes, this small upgrade can improve the bottom line by more than $1 million, thanks to increased productivity and reduced waste. It’s clear that these economic advantages completely overwhelm the conventional wisdom of purchasing a valve on the basis of only the initial purchase price.
Secondly, conventional wisdom has always maintained that improved process optimization comes from better control room instrumentation. However, the test data in Figure 1 shows that under the same instrumentation, the valve’s dynamic performance can affect loop performance significantly. Little can be gained by spending a lot of money for a sophisticated control instrumentation system capable of performing to 0.5 percent accuracy, if the accuracy of the ontrol valve is only five percent.
Valve gain and characteristics
Conventional thinking about sizing focuses on whether a valve can pass the required flow under all service conditions. Because of uncertainties, it’s common for the specifying engineer to provide some additional margin of flow. And it’s not uncommon for even more margin to be added during subsequent equipment reviews. In other situations, particularly with butterfly valves, it’s common to specify line-sized valves. As a result, the majority of control valves in service today are grossly oversized. Until recently, no one understood how much this affects valve gain, which, in turn, influences process variability.
Valve gain is the flow rate change with respect to valve travel. Under static test conditions of constant pressure drop, a curve showing steady-state flow, plotted as a function of valve travel, would be called an “inherent valve characteristic.” It represents how a particular valve operates when it’s subjected to constant pressure drop conditions. The three inherent valve characteristics are quick opening, linear and equal percentage.
Values with quick openings have high gain at the lower travels, with each travel increment producing large increases in flow. As the valve opening increases, the gain decreases proportionately. Valves with linear characteristics have constant gain throughout their complete valve travel. Meanwhile, valves with equal percentage characteristics show a corresponding increase in valve travel and flow.
We choose a valve based on an inherent characteristic to compensate for gain changes in other parts of the loop. Without a constant loop gain, it becomes difficult to realize optimum control, and process variability will suffer.
However, when a control valve is installed and subjected to a pressure drop that varies with process conditions, it is unlikely the valve will exhibit any of its inherent flow characteristics. In fact, it would be inappropriate for the valve’s installed characteristic to be anything other than linear. The installed gain of the entire loop remains constant.
Limitations in available hardware make it difficult to achieve a constant loop gain of 1.0 throughout complete valve travel. Most process control experts agree that a loop gain variation of ± 2.0 is acceptable. Thus, a valves’s loop gain of 0.5 to 2.0 is defined as an acceptable control range.
Note that neither valve has the desired linear installed characteristic. However, the globe valve comes a lot closer to being linear than does the butterfly valve. From the lower curve, we see that the globe valve has a larger control range (55 percent) than the butterfly valve (20 percent). In the valve performance hierarchy, globe valves demonstrate the widest control range, followed by V-notch ball valves and eccentric disk-style valves. Butterfly valves typically have the narrowest and are generally suited only for fixed-load applications.
Regardless of inherent characteristic or design style, a severely oversized valve acts like a quick-opening valve, with high installed gain in the lower lift regions. Most importantly, process optimization requires that a valve’s style and size remain within the control range over the widest possible operating range. Selection of an inappropriate style or size results in poor dynamic loop performance.
Valve style considerations
Selecting the right valve for an application can be made easier by first reviewing the four basic styles of throttling-control valves: cage-style globe valves, ball valves, eccentric disk valves and butterfly valves.